Structures of riboswitch RNA reaction states by mix-and-inject XFEL serial crystallography.

Riboswitches are structural RNA elements that are generally located in the 5' untranslated region of messenger RNA. During regulation of gene expression, ligand binding to the aptamer domain of a riboswitch triggers a signal to the downstream expression platform. A complete understanding of the structural basis of this mechanism requires the ability to study structural changes over time. Here we use femtosecond X-ray free electron laser (XFEL) pulses to obtain structural measurements from crystals so small that diffusion of a ligand can be timed to initiate a reaction before diffraction. We demonstrate this approach by determining four structures of the adenine riboswitch aptamer domain during the course of a reaction, involving two unbound apo structures, one ligand-bound intermediate, and the final ligand-bound conformation. These structures support a reaction mechanism model with at least four states and illustrate the structural basis of signal transmission. The three-way junction and the P1 switch helix of the two apo conformers are notably different from those in the ligand-bound conformation. Our time-resolved crystallographic measurements with a 10-second delay captured the structure of an intermediate with changes in the binding pocket that accommodate the ligand. With at least a 10-minute delay, the RNA molecules were fully converted to the ligand-bound state, in which the substantial conformational changes resulted in conversion of the space group. Such notable changes in crystallo highlight the important opportunities that micro- and nanocrystals may offer in these and similar time-resolved diffraction studies. Together, these results demonstrate the potential of 'mix-and-inject' time-resolved serial crystallography to study biochemically important interactions between biomacromolecules and ligands, including those that involve large conformational changes.

Crystal structure of a conserved ribosomal protein-RNA complex.

The structure of a highly conserved complex between a 58-nucleotide domain of large subunit ribosomal RNA and the RNA-binding domain of ribosomal protein L11 has been solved at 2.8 angstrom resolution. It reveals a precisely folded RNA structure that is stabilized by extensive tertiary contacts and contains an unusually large core of stacked bases. A bulge loop base from one hairpin of the RNA is intercalated into the distorted major groove of another helix; the protein locks this tertiary interaction into place by binding to the intercalated base from the minor groove side. This direct interaction with a key ribosomal RNA tertiary interaction suggests that part of the role of L11 is to stabilize an unusual RNA fold within the ribosome.

How do proteins recognize specific RNA sites? New clues from autogenously regulated ribosomal proteins.

Some ribosomal proteins which bind specifically to ribosomal RNA also act as translational repressors and recognize their encoding messenger RNAs. The messenger- and ribosomal-RNA binding sites for four of these proteins are now well defined, and striking similarities in primary and secondary structure are apparent in most cases. These 'consensus' structures are useful clues to the features proteins use to recognize specific RNAs.

Strategies for RNA folding.

RNAs are surprisingly adept at folding into specific shapes capable of ligand recognition and catalysis. Thermodynamic analysis of the unfolding of several different RNAs suggests that there are at least three strategies an RNA might use to achieve a very stable and compactly folded structure: hydrogen bonding between irregular complementary surfaces (as in transfer RNA tertiary structure); monovalent and divalent lons bound to specific sites (as found in a ribosomal RNA fragment) and pseudoknot folds (exemplified by a messenger RNA fragment with extensive non-canonical structure).

RNA structure.

New information concerning RNA structure is accumulating at an ever increasing rate-from short helices with mismatched bases of 5S rRNA and complex RNA aptamers. The importance of recurring structural motifs, ion binding, and the kinetics and energetics of folding in RNA structure and function is now being recognized and addressed.

The RNA-binding domain of ribosomal protein L11 recognizes an rRNA tertiary structure stabilized by both thiostrepton and magnesium ion.

Antibiotics that inhibit ribosomal function may do so by one of several mechanisms, including the induction of incorrect RNA folding or prevention of protein and/or RNA conformational transitions. Thiostrepton, which binds to the 'GTPase center' of the large subunit, has been postulated to prevent conformational changes in either the L11 protein or rRNA to which it binds. Scintillation proximity assays designed to look at the binding of the L11 C-terminal RNA-binding domain to a 23S ribosomal RNA (rRNA) fragment, as well as the ability of thiostrepton to induce that binding, were used to demonstrate the role of Mg(2+), L11 and thio-strepton in the formation and maintenance of the rRNA fragment tertiary structure. Experiments using these assays with both an Escherichia coli rRNA fragment and a thermostable variant of that RNA show that Mg(2+), L11 and thiostrepton all induce the RNA to fold to an essentially identical tertiary structure.

Dissection of the 16S rRNA binding site for ribosomal protein S4.

The ribosomal protein S4 from Escherichia coli is essential for initiation of assembly of 30S ribosomal subunits. We have undertaken the identification of specific features required in the 16S rRNA for S4 recognition by synthesizing mutants bearing deletions within a 460 nucleotide region which contains the minimum S4 binding site. We made a set of large nested deletions in a subdomain of the molecule, as well as individual deletions of nine hairpins, and used a nitrocellulose filter binding assay to calculate association constants. Some small hairpins can be eliminated with only minor effects on S4 recognition, while three hairpins scattered throughout the domain (76-90, 376-389 and 456-476) are essential for specific interaction. The loop sequence of hairpin 456-476 is important for S4 binding, and may be directly recognized by the protein. Some of the essential features are in phylogenetically variable regions; consistent with this, Mycoplasma capricolum rRNA is only weakly recognized by S4, and no specific binding to Xenopus laevis rRNA can be detected.

Themes in RNA-protein recognition.

Atomic resolution structures are now available for more than 20 complexes of proteins with specific RNAs. This review examines two main themes that appear in this set of structures. A "groove binder" class of proteins places a protein structure (alpha-helix, 310-helix, beta-ribbon, or irregular loop) in the groove of an RNA helix, recognizing both the specific sequence of bases and the shape or dimensions of the groove, which are sometimes distorted from the normal A-form. A second class of proteins uses beta-sheet surfaces to create pockets that examine single-stranded RNA bases. Some of these proteins recognize completely unstructured RNA, and in others RNA secondary structure indirectly promotes binding by constraining bases in an appropriate orientation. Thermodynamic studies have shown that binding specificity is generally a function of several factors, including base-specific hydrogen bonds, non-polar contacts, and mutual accommodation of the protein and RNA-binding surfaces. The recognition strategies and structural frameworks used by RNA binding proteins are not exotically different from those employed by DNA-binding proteins, suggesting that the two kinds of nucleic acid-binding proteins have not evolved independently.

Contributions of basic residues to ribosomal protein L11 recognition of RNA.

The C-terminal domain of ribosomal protein L11, L11-C76, binds in the distorted minor groove of a helix within a 58 nucleotide domain of 23 S rRNA. To study the electrostatic component of RNA recognition in this protein, arginine and lysine residues have been individually mutated to alanine or methionine residues at the nine sequence positions that are conserved as basic residues among bacterial L11 homologs. In measurements of the salt dependence of RNA-binding, five of these mutants have a reduced value of - partial differentiallog(K(obs))/ partial differentiallog[KCl] as compared to the parent L11-C76 sequence, indicating that these residues interact with the RNA electrostatic field. These five residues are located at the perimeter of the RNA-binding surface of the protein; all five of them form salt bridges with phosphates in the crystal structure of the complex. A sixth residue, Lys47, was found to make an electrostatic contribution to binding when measurements were made at pH 6.0, but not at pH 7.0; its pK in the free protein must be <6.5. The unusual behavior of Lys47 is explained by its burial in the hydrophobic core of the free protein, and unburial in the RNA-bound protein, where it forms a salt bridge with a phosphate. The contributions of these six residues to the electrostatic component of binding are not additive; thus the magnitude of the salt dependence cannot be used to count the number of ionic interactions in this complex. By interacting with irregular features of the RNA backbone, including an S-turn, these basic residues contribute to the specificity of L11 for its target site.

S4-16 S ribosomal RNA complex. Binding constant measurements and specific recognition of a 460-nucleotide region.

The region of the Escherichia coli 16 S ribosomal RNA recognized by the ribosomal protein S4 has been defined by assaying a set of 13 16 S rRNA fragments for S4 binding. The fragments were prepared by transcription in vitro, and binding constants were measured in three ways: retention of labeled RNA fragments on nitrocellulose filters by S4; co-sedimentation of labeled S4 with RNA fragments in sucrose gradients; and the distribution of labeled S4 between two RNAs of different sizes in a sucrose gradient. All three methods gave similar relative binding strengths for a variety of 16 S rRNA and non-specific (23 S rRNA) sequences, with the exception of two of the largest 16 S rRNA fragments; these gave smaller association constants in the filter retention assay than in the other methods. We found that specific complexes of S4 with these larger RNAs do not bind well to filters, leaving non-specific complexes to dominate the assay. Specific complexes with RNAs less than or equal to 891 nucleotides were retained efficiently by S4 on filters, and gave reliable binding constants. All 16 S rRNA fragments containing nucleotides 39 to 500 bound S4 with the same affinity as intact 16 S rRNA, while all fragments with endpoints within 39 to 500 bound at least tenfold more weakly. This sequence must be able to fold independently of the rest of the rRNA. Comparison of this minimal 462-nucleotide S4 binding site with S4 footprinting results suggests that S4 binding might alter the conformations of RNA neighboring the 39 to 500 region in the intact 16 S rRNA. Specific S4-rRNA binding is not sensitive to KCl concentration, but a more normal salt dependence is seen in K2SO4 (delta logK/delta log[K+] approximately -3.3). This duplicates the behavior of the specific S4-alpha mRNA translational repression complex, arguing that S4 recognizes both the mRNA and rRNA substrates by the same mechanism. Mg2+ is not required to form the specific rRNA complex, at least under conditions which stabilize RNA structure (0.35 M-KCl, 5 degrees C).

Stabilization of RNA tertiary structure by monovalent cations.

The effects of monovalent cations (Li(+), Na(+), K(+), Rb(+), Cs(+), and NH4(+)) on the thermal stability of RNA tertiary structure were investigated by UV melting. We show that with the RNA used here (nucleotides 1051-1108 of Escherichia coli 23 S rRNA with four base substitutions), monovalent cations and Mg(2+) compete in stabilizing the RNA tertiary structure, and that the competition takes place between two boundaries: one where Mg(2+) concentration is zero and the other where it is maximally stabilizing ("saturating"). The pattern of competition is the same for all monovalent cations and depends on the cation's ability to displace Mg(2+) from the RNA, its ability to stabilize tertiary structure in the absence of Mg(2+), and its ability to stabilize tertiary structure at saturating Mg(2+) concentrations. The stabilizing ability of a monovalent cation depends on its unhydrated ionic radius, and at a low monovalent cation concentration and saturating Mg(2+), there is a (calculated) net release of a single monovalent cation/RNA molecule when tertiary structure is denatured. The implications are that under these conditions there is at least one binding site for monovalent cations on the RNA, the site is specifically associated with formation of stable tertiary structure, K(+) is the most effective of the tested cations, and Mg(2+) appears ineffective at this site. At high ionic strength, and in the absence of Mg(2+), stabilization of tertiary structure is still monovalent-cation specific and ionic-radius dependent, but a larger number of cations ( approximately eight) are released upon RNA tertiary structure denaturation, and NH(4)(+) appears to be the most effective cation in stabilizing tertiary structure under these conditions. In the majority of the experiments, methanol was added as a cosolvent to the buffer. Its use allowed the examination of the behavior of monovalent ions under conditions where their effects would otherwise have been too weak to be observed. Methanol stabilizes tertiary but not secondary structure of the RNA. There was no evidence that it either causes qualitative changes in cation-binding properties of the RNA or a change in the pattern of monovalent cation/Mg(2+) competition.

Stabilization of a ribosomal RNA tertiary structure by ribosomal protein L11.

Interactions between ribosomal protein L11 and a domain of large subunit rRNA have been highly conserved and are essential for efficient protein synthesis. To study the effects of L11 on rRNA folding, a homolog of the Escherichia coli L11 gene has been amplified from Bacillus stearothermophilus DNA and cloned into a phage T7 polymerase-based expression system. The expressed protein is 93% homologous to the L11 homolog from Bacillus subtilis, denatures at temperatures above 72 degrees C, and has nearly identical rRNA binding properties as the Escherichia coli L11 in terms of RNA affinity constants and their dependences on temperature, Mg2+ concentration, monovalent cation, and RNA mutations. Mg2+ and NH4+ are specifically bound by the RNA-protein complex, with apparent ion-RNA affinities of 1.6 mM-1 and 19 M-1, respectively, at 0 degree C. The effect of the thermostable L11 on the unfolding of a 60 nucleotide rRNA fragment containing its binding domain has been examined in melting experiments. The lowest temperature RNA transition, which is attributed to tertiary structure unfolding, is stabilized by approximately 25 degrees C, and the interaction has an intrinsic enthalpy of approximately 13 kcal/mol. The thermal stability of the protein-RNA complex is enhanced by increasing Mg2+ concentration and by NH4+ relative to Na+. Thus L11, NH4+, and Mg2+ all bind and stabilize the same rRNA tertiary interactions, which are conserved and presumably important for ribosome function.

Affinities and selectivities of divalent cation binding sites within an RNA tertiary structure.

A 58 nucleotide fragment of Escherichia coli large subunit ribosomal RNA, nucleotides 1051 to 1108, adopts a specific tertiary structure normally requiring both monovalent (NH4+ or K+) and divalent (Mg2+) ions to fold; this ion-dependent structure is a prerequisite for recognition by ribosomal protein L11. Melting experiments have been used to show that a sequence variant of this fragment, GACG RNA, is able to adopt a stable tertiary structure in the presence of 1.6 M NH4Cl and absence of divalent ions. The similarity of this high-salt structure to the tertiary structure formed under more typical salt conditions (0.1 M NH4Cl and several mM MgCl2) was shown by its following properties: (i) an unusual ratio of hyperchromicity at 260 nm and 280 nm upon unfolding, (ii) selectivity for NH4+ over K+ or Na+, (iii) stabilization by L11 protein, and (iv) further stabilization by added Mg2+. Delocalized electrostatic interactions of divalent ions with nucleic acids should be very weak in the presence of >1 M monovalent salt; thus stabilization of the tertiary structure by low (<1 mM) Mg2+ concentrations in these high-salt conditions suggests that Mg2+ binds at specific site(s). GACG RNA tertiary structure unfolding in 1.6 M NH4Cl (Tm approximately 39 degrees C) is distinct from melting of the secondary structure (centered at approximately 72 degrees C), and it has been possible to calculate the free energy of tertiary structure stabilization upon addition of various divalent cations. From these binding free energies, ion-RNA binding isotherms for Mn2+, Mg2+, Ca2+, Sr2+ and Ba2+ have been obtained. All of these ions bind at two sites: one site favors Mg2+ and Ba2+ and discriminates against Ca2+, while the other site favors binding of smaller ions over larger ones (Mg2+ >Ca2+ >Sr2+ >Ba2+). Weak cooperative or anticooperative interactions between the sites, also dependent on ion radius, may also be taking place.

Mg(2+) binding to tRNA revisited: the nonlinear Poisson-Boltzmann model.

Our current understanding of Mg(2+) binding to RNA, in both thermodynamic and structural terms, is largely based on classical studies of transfer RNAs. Based on these studies, it is clear that magnesium ions are crucial for stabilizing the folded structure of tRNA. We present here a rigorous theoretical model based on the nonlinear Poisson-Boltzmann (NLPB) equation for understanding Mg(2+) binding to yeast tRNA(Phe). We use this model to interpret a variety of experimental Mg(2+) binding data. In particular, we find that the NLPB equation provides a remarkably accurate description of both the overall stoichiometry and the free energy of Mg(2+) binding to yeast tRNA(Phe) without any fitted parameters. In addition, the model accurately describes the interaction of Mg(2+) with localized regions of the RNA as determined by the pK(a) shift of differently bound fluorophores. In each case, we find that the model also reproduces the univalent salt-dependence and the anticooperativity of Mg(2+) binding. Our results lead us to a thermodynamic description of Mg(2+) binding to yeast tRNA(Phe) based on the NLPB equation. In this model, Mg(2+) binding is simply explained by an ensemble of ions distributed according to a Boltzmann weighted average of the mean electrostatic potential around the RNA. It appears that the entire ensemble of electrostatically bound ions superficially mimics a few strongly coordinated ions. In this regard, we find that Mg(2+) stabilizes the tertiary structure of yeast tRNA(Phe) in part by accumulating in regions of high negative electrostatic potential. These regions of Mg(2+) localization correspond to bound ions that are observed in the X-ray crystallographic structures of yeast tRNA(Phe). Based on our results and the available thermodynamic data, there is no evidence that specifically coordinated Mg ions have a significant role in stabilizing the native tertiary structure of yeast tRNA(Phe) in solution.

Thermodynamics of RNA folding in a conserved ribosomal RNA domain.

A small, 58 nt domain of the large subunit ribosomal RNA (Escherichia coli sequence 1051 to 1108) is a highly conserved junction of three helices whose secondary structure has been established by phylogenetic comparisons. To detect any contributions of additional tertiary interactions, the thermal denaturation of the rRNA domain was followed by either UV hyperchromicity or calorimetry in buffers containing a wide range of Mg2+ concentrations. Several smaller fragments corresponding to two different hairpin stem-loop structures within the domain were also synthesized and melted for comparison with the larger molecule. A model of the secondary structure unfolding was devised, based on measured enthalpies and melting temperatures of the component hairpins and tabulated parameters of base-pair stacking and loop closure. The model closely simulates the observed melting data when three additional factors are included: two parameters to account for coaxial stackings within a junction of helices, and a set of undefined "tertiary" interactions that unfolds before the secondary structure and is preferentially stabilized by Mg2+. A critical feature of this model is a conserved pair, U1082/A1086, that is within the junction loop and hypothesized to stack with an adjacent helix. The model correctly predicts the effects of disrupting this pair in a U1086 sequence variant. Although the set of "tertiary" interactions contributes a significant fraction of the RNA unfolding enthalpy (delta H approximately 25 kcal/mol, out of 180 kcal/mol total), its overall stability is marginal at 37 degrees C.

Bases defining an ammonium and magnesium ion-dependent tertiary structure within the large subunit ribosomal RNA.

A 58 nucleotide RNA derived from a highly conserved domain of the large subunit ribosomal RNA (Escherichia coli 1051 to 1108) has a set of tertiary interactions that is stabilized by NH4+ and Mg2+ in preference to other ions. We have mapped the nucleotides contributing to this structure by examining the thermal denaturation of 25 sequence variants. Where necessary compensatory mutations were made to preserve the phylogenetically conserved secondary structure. Substitutions of bases or base-pairs at eight positions specifically eliminate the ion-dependent tertiary structure without affecting the secondary structure stability; most of these positions are conserved among all large subunit RNA sequences. At two positions, substitutions of bases found in other organisms stabilize the E. coli tertiary structure by substantial amounts (delta G 37 degrees becomes more favorable by -1.8 to -4.5 kcal/mol). One of these variants disrupts a potential A.U base-pair within a helix, suggesting that the tertiary structure competes with alternative structures. The results show that this rRNA domain contains an extensive, highly conserved, and very stable set of tertiary interactions. The sequence in E. coli, and probably most other organisms, has not evolved to maximum stability. It is possible that natural selection has "tuned" the tertiary structure to an optimum stability, perhaps because the structure must open and close during the ribosome cycle.

Thermodynamics of folding a pseudoknotted mRNA fragment.

A sequence in the leader and first gene of the Escherichia coli alpha mRNA folds into a complex pseudoknot structure that is required for binding of a translational repressor. The thermal denaturation of a 112 nt RNA containing this structure has been followed by calorimetry and UV hyperchromicity. To determine the partially folded intermediates in unfolding, the denaturation of 13 mutants and of several fragments with successive deletions of helices were investigated as well. An unfolding pathway with seven states is proposed as the simplest mechanism that accounts for the data, and has several implications. (1) The lowest temperature transition appears only in the presence of moderate concentrations of Mg2+ or high concentrations of K+ (delta H approximately 45 kcal/mol), and is the unfolding of tertiary structures, rather than secondary structure. Under some conditions it is destabilized by increasing salt concentration. (2) Two of the intermediates unfolding at higher temperature must have non-canonical or tertiary interactions in addition to the known secondary structure. (3) Two alternative structures compete for formation of the complete pseudoknot, and form as the pseudoknot unfolds. Thus structures not present in the completely folded pseudoknot affect the overall thermodynamics, and probably the kinetics, of unfolding. (4) Approximately 16 kcal/mol of free energy is required to completely expose the coding region to ribosomes at 37 degrees C, though approximately 6.5 kcal/mol is regained by refolding of upstream regions after the pseudoknot is unfolded. The substantial energy needed to unfold the pseudoknot may affect the rate of translation from this ribosome binding site. A simple model of RNA folding in which an optimum secondary structure forms first, followed by tertiary interactions that further stabilize the secondary structure, does not hold in this RNA.

S4-alpha mRNA translation regulation complex. II. Secondary structures of the RNA regulatory site in the presence and absence of S4.

The secondary structure of the Escherichia coli alpha mRNA leader sequence has been determined using nucleases specific for single- or double-stranded RNA. Three different length alpha RNA fragments were studied at 0 degrees C and 37 degrees C. A very stable eight base-pair helix forms upstream from the ribosome initiation site, defining a 29 base loop. There is evidence for base-pairing between nucleotides within this loop and for a "pseudo-knot" interaction of some loop bases with nucleotides just 3' to the initiation codon, forming a region of complex structure. A weak helix also pairs sequences near the 5' terminus of the alpha mRNA with bases near the Shine-Dalgarno sequence. Affinity constants for the translational repressor S4 binding different length alpha mRNA fragments indicate that most of the S4 recognition features must be contained within the main helix and hairpin regions. Binding of S4 to the alpha mRNA alters the structure of the 29 base hairpin region, and probably melts the weak pairing between the 5' and 3' termini of the leader. The pseudo-knot structure and the conformational changes associated with it provide a link between the structures of the S4 binding site and the ribosome binding site. The alpha mRNA may therefore play an active role in mediating translational repression.

The interpretation of Mg(2+) binding isotherms for nucleic acids using Poisson-Boltzmann theory.

Magnesium ions play a crucial role in the structural integrity and biological activity of nucleic acids. Experimental thermodynamic descriptions of Mg(2+) interactions with nucleic acids in solution have generally relied on the analyses of binding polynomials to estimate the energetic contributions of diffuse and site-bound ions. However, since ion binding is dominated by long-range electrostatic forces, such models provide only a phenomenological description of the experimental Mg(2+) binding data and provide little insight into the actual mechanism of the binding equilibria. Here, we present a rigorous theoretical framework based on the non-linear Poisson-Boltzmann (NLPB) equation for understanding diffuse ion interactions that can be used to interpret experimental Mg(2+) binding isotherms. As intuitively expected, in the NLPB model binding is simply the total accumulation of the ion around the nucleic acid. Comparing the experimental data to the calculated curves shows that the NLPB equation provides a remarkably accurate description of Mg(2+) binding to linear polynucleotides like DNA and poly(A x U) without any fitted parameters. In particular, the NLPB model explains two general features of magnesium binding; the strong dependence on univalent salt concentration, and its substantial anticooperativity. Each of these effects can be explained by changes in the Mg(2+) distribution around the polyion under different solution conditions. In order to more fully understand these different aspects of magnesium binding, the free energy of Mg(2+) binding, DeltaGMg, is calculated and partitioned into several salt-dependent contributions: the change in the electrostatic interaction free energy of the charges, DeltaDeltaGE.D (including Mg(2+)-phosphate, Mg(2+)-Mg(2+), Mg(2+)-Na(+), Na(+)-Na(+), Na(+)-phosphate interactions, and similar contributions for Cl(-)) and the cratic free energies of (re)organizing the MgCl2 and NaCl atmospheres, DeltaG(Mg)org and DeltaDeltaG(Na)org, respectively. For the systems studied here, DeltaGMg is strongly influenced by entropic free energy changes in the distributions of both NaCl and MgCl2, DeltaG(Mg)org and DeltaDeltaG(Na)org. From this analysis, we also raise the possibility that coions added with the magnesium salt might play an important role in the overall stability of nucleic acids under some conditions.